Certain DNSSEC aspects of the DNS protocol (in RFC 4033, 4034, 4035, 6840, and related RFCs) allow remote attackers to cause a denial of service (CPU consumption) via one or more DNSSEC responses, aka the "KeyTrap" issue. One of the concerns is that, when there is a zone with many DNSKEY and RRSIG records, the protocol specification implies that an algorithm must evaluate all combinations of DNSKEY and RRSIG records.

A flaw in the networking code handling DNS-over-TLS queries may cause `named` to terminate unexpectedly due to an assertion failure. This happens when internal data structures are incorrectly reused under significant DNS-over-TLS query load. This issue affects BIND 9 versions 9.18.0 through 9.18.18 and 9.18.11-S1 through 9.18.18-S1.

The code that processes control channel messages sent to `named` calls certain functions recursively during packet parsing. Recursion depth is only limited by the maximum accepted packet size; depending on the environment, this may cause the packet-parsing code to run out of available stack memory, causing `named` to terminate unexpectedly. Since each incoming control channel message is fully parsed before its contents are authenticated, exploiting this flaw does not require the attacker to hold a valid RNDC key; only network access to the control channel's configured TCP port is necessary. This issue affects BIND 9 versions 9.2.0 through 9.16.43, 9.18.0 through 9.18.18, 9.19.0 through 9.19.16, 9.9.3-S1 through 9.16.43-S1, and 9.18.0-S1 through 9.18.18-S1.

Every `named` instance configured to run as a recursive resolver maintains a cache database holding the responses to the queries it has recently sent to authoritative servers. The size limit for that cache database can be configured using the `max-cache-size` statement in the configuration file; it defaults to 90% of the total amount of memory available on the host. When the size of the cache reaches 7/8 of the configured limit, a cache-cleaning algorithm starts to remove expired and/or least-recently used RRsets from the cache, to keep memory use below the configured limit. It has been discovered that the effectiveness of the cache-cleaning algorithm used in `named` can be severely diminished by querying the resolver for specific RRsets in a certain order, effectively allowing the configured `max-cache-size` limit to be significantly exceeded. This issue affects BIND 9 versions 9.11.0 through 9.16.41, 9.18.0 through 9.18.15, 9.19.0 through 9.19.13, 9.11.3-S1 through 9.16.41-S1, and 9.18.11-S1 through 9.18.15-S1.

This issue can affect BIND 9 resolvers with `stale-answer-enable yes;` that also make use of the option `stale-answer-client-timeout`, configured with a value greater than zero. If the resolver receives many queries that require recursion, there will be a corresponding increase in the number of clients that are waiting for recursion to complete. If there are sufficient clients already waiting when a new client query is received so that it is necessary to SERVFAIL the longest waiting client (see BIND 9 ARM `recursive-clients` limit and soft quota), then it is possible for a race to occur between providing a stale answer to this older client and sending an early timeout SERVFAIL, which may cause an assertion failure. This issue affects BIND 9 versions 9.16.12 through 9.16.36, 9.18.0 through 9.18.10, 9.19.0 through 9.19.8, and 9.16.12-S1 through 9.16.36-S1.

BIND 9 resolver can crash when stale cache and stale answers are enabled, option `stale-answer-client-timeout` is set to a positive integer, and the resolver receives an RRSIG query. This issue affects BIND 9 versions 9.16.12 through 9.16.36, 9.18.0 through 9.18.10, 9.19.0 through 9.19.8, and 9.16.12-S1 through 9.16.36-S1.

Sending a flood of dynamic DNS updates may cause `named` to allocate large amounts of memory. This, in turn, may cause `named` to exit due to a lack of free memory. We are not aware of any cases where this has been exploited. Memory is allocated prior to the checking of access permissions (ACLs) and is retained during the processing of a dynamic update from a client whose access credentials are accepted. Memory allocated to clients that are not permitted to send updates is released immediately upon rejection. The scope of this vulnerability is limited therefore to trusted clients who are permitted to make dynamic zone changes. If a dynamic update is REFUSED, memory will be released again very quickly. Therefore it is only likely to be possible to degrade or stop `named` by sending a flood of unaccepted dynamic updates comparable in magnitude to a query flood intended to achieve the same detrimental outcome. BIND 9.11 and earlier branches are also affected, but through exhaustion of internal resources rather than memory constraints. This may reduce performance but should not be a significant problem for most servers. Therefore we don't intend to address this for BIND versions prior to BIND 9.16. This issue affects BIND 9 versions 9.16.0 through 9.16.36, 9.18.0 through 9.18.10, 9.19.0 through 9.19.8, and 9.16.8-S1 through 9.16.36-S1.

By sending specific queries to the resolver, an attacker can cause named to crash.

An attacker can leverage this flaw to gradually erode available memory to the point where named crashes for lack of resources. Upon restart the attacker would have to begin again, but nevertheless there is the potential to deny service.

The underlying bug might cause read past end of the buffer and either read memory it should not read, or crash the process.

By flooding the target resolver with queries exploiting this flaw an attacker can significantly impair the resolver's performance, effectively denying legitimate clients access to the DNS resolution service.

There had existed in one of the ISC BIND libraries a bug in a function that was used by dhcpd when operating in DHCPv6 mode. There was also a bug in dhcpd relating to the use of this function per its documentation, but the bug in the library function prevented this from causing any harm. All releases of dhcpd from ISC contain copies of this, and other, BIND libraries in combinations that have been tested prior to release and are known to not present issues like this. Some third-party packagers of ISC software have modified the dhcpd source, BIND source, or version matchup in ways that create the crash potential. Based on reports available to ISC, the crash probability is large and no analysis has been done on how, or even if, the probability can be manipulated by an attacker. Affects: Builds of dhcpd versions prior to version 4.4.1 when using BIND versions 9.11.2 or later, or BIND versions with specific bug fixes backported to them. ISC does not have access to comprehensive version lists for all repackagings of dhcpd that are vulnerable. In particular, builds from other vendors may also be affected. Operators are advised to consult their vendor documentation.